High-Frequency Voltage-Injection Methods and Observer Design for Initial Position Detection of Permanent Magnet Synchronous Machines

Xinhai Jin, Ronggang Ni, Wei Chen, Frede Blaabjerg, Dianguo Xu

Research output: Contribution to journalJournal articleResearchpeer-review

53 Citations (Scopus)
820 Downloads (Pure)

Abstract

The information of the initial rotor position is essential for smooth start up and robust control of Permanent Magnet Synchronous Machines (PMSMs). RoTating Voltage Injection (RTVI) methods in the stationary reference frame have been commonly adopted to detect the initial rotor position at standstill without any position sensors. However, alt-hough the Pulsating sqUare-wave Voltage Injection (PUVI) method performs better in estimation time and accuracy, it is rarely used because the estimation result may converge to the q-axis. In this paper, this fault convergence is avoided by modifying the initial states of the position observer, and the PUVI method can finally be used for robust initial rotor position detection. Modified signal processing techniques are proposed for both RTVI and PUVI methods for better implementations in fixed point processors and easier observer gain designs. Detailed comparisons between these two methods are provided. Furthermore, two position estimation observers, i.e. the Proportional-Integral (PI) observer and the Extended State Observer (ESO) are compared, and their parameter tuning methods are studied as well. Both simulation and experimental results are provided for verifications.
Original languageEnglish
JournalIEEE Transactions on Power Electronics
Volume33
Issue number9
Pages (from-to)7971-7979
Number of pages9
ISSN0885-8993
DOIs
Publication statusPublished - Sept 2018

Keywords

  • Initial rotor position detection
  • Observer
  • Permanent Magnet Synchronous Machine (PMSM)
  • Position sensorless control
  • Square-wave voltage injection

Fingerprint

Dive into the research topics of 'High-Frequency Voltage-Injection Methods and Observer Design for Initial Position Detection of Permanent Magnet Synchronous Machines'. Together they form a unique fingerprint.

Cite this